Validation of a fast LC-MS/MS method for the analysis of bisphenol A (BPA) and urethane dimethacrylate (UDMA) in eluates of dental polymer-based materials

Abstract

Polymer-based materials are widely used in dentistry. However, concerns exist regarding their biocompatibility because monomers such as urethane dimethacrylate (UDMA) and bisphenol A (BPA) may leach from the materials and affect the immune and endocrine systems. Since BPA and UDMA are biologically active even at low concentrations, validated and highly sensitive analytical methods are required. We developed a fast LC-MS/MS method enabling simultaneous detection of BPA and UDMA in artificial saliva using a short column and polarity switching. The method was validated for selectivity, specificity, carry-over, sensitivity, accuracy, precision, recovery, and matrix effects. A water–methanol gradient elution with post-column infusion of 6 mM NH₄F was used to enhance sensitivity. Lower limits of quantification (LLOQ) for UDMA and BPA were of 10pg/ml and 30pg/ml, respectively. The method was used to analyze eluates of thermoformed and 3D-printed polymer-based materials, which released very low amounts of BPA and UDMA. In eluates collected after 1 day, the thermoformed material released more BPA (154.6 ± 128.7 pg/mL) and UDMA (154.8 ± 24.0 pg/mL) than the 3D-printed resin (BPA: 30.7 ± 16.4 pg/mL, UDMA: under LLOQ). Despite the high sensitivity, all values were lower than LLOQ in eluates collected after 1 week. The method provides a fast and reliable tool for the simultaneous quantification of BPA and UDMA in the eluates of dental polymer-based materials, enabling a more accurate assessment of the safety of these materials.

Introduction

Polymer-based materials are widely used in dentistry. Among these, resin-based composites are the most commonly used, especially as amalgam is being phased out due to environmental concerns and potential health effects^1,2. Other applications include resin-based sealants in pediatric dentistry, resin-modified glass ionomers and adhesives in restorative dentistry, splints, temporary crowns and dentures in prosthodontics, as well as clear aligners or palatal plates in orthodontics. Depending on their specific formulation and intended clinical use, polymer-based materials may offer advantages such as esthetic appearance, light-curing capability, ease of manipulation, or flexibility in design and manufacturing.

Common monomers used in dental resins are methyl methacrylate (MMA), 2-hydroxyethyl methacrylate (HEMA) or dimethacrylates such as triethylene glycol dimethacrylate (TEGDMA), urethane dimethacrylate (UDMA), bisphenol A glycidyl dimethacrylate (Bis-GMA), and its ethoxylated derivative (Bis-EMA) (Fig. 1^3,4;). The monomer constitution of dental resins is tailored to the specific application and should be selected based on the required mechanical and optical properties, viscosity, and polymerization shrinkage⁵.

Fig. 1.

Chemical structures of the most common monomers in dental resins.

However, despite functional benefits of dental polymer-based materials, concerns have been raised about their biocompatibility. This is particularly due to the potential release of residual monomers into the oral environment, which may occur initially as a result of incomplete polymerization and later due to material degradation over time¹.

Methacrylate monomers can exhibit various adverse effects, both to dental personnel and patients on local or systemic level¹. Besides the allergenic potential of low-molecular weight monomers like HEMA or TEGDMA⁶, dental methacrylate monomers were documented to be cytotoxic^7,8 with toxicity decreasing in the order BisGMA > UDMA > TEGDMA > HEMA⁹. Along with other ingredients of dental materials, such as polymerization initiators¹⁰, methacrylates were also reported to be genotoxic and mutagenic^11,12. In addition, recent studies indicated that their immunomodulatory properties may disturb the delicate balance of immune cells and lead to a higher release of IL-8³, which is a cytokine involved in inflammatory response.

Last but not least, materials containing monomers based on bisphenol A (BPA), such as Bis-GMA and Bis-EMA, can release BPA as an contaminant or degradation product¹³, potentially disrupting the endocrine system^2,14. Additionally, low-dose effects of BPA on immune system are now widely discussed¹⁵. Besides its use in dental materials, BPA has widespread applications as a component of epoxy resins and polycarbonate plastics, which are found in food contact materials (e.g., cans, bottles) as well as in other products such as thermal paper¹⁶. Concerns about BPA release have been further raised by the European Food Safety Authority (EFSA), which recommended to lower the tolerable daily intake of BPA 20,000 times — from 4 µg/kg body weight (bw) to 0.2 ng/kg bw¹⁷.

With the increasing adoption of digital workflows in dentistry, three-dimensional (3D) printing has introduced a new generation of polymer-based materials. While these materials offer distinct advantages in terms of customization and fabrication speed, their biocompatibility remains a matter of active investigation¹⁸. For instance, clear aligners are gaining popularity due to their improved aesthetics and greater comfort compared to traditional fixed orthodontic appliances¹⁹. However, while 3D-printed aligners exhibit potential to replace thermoformed aligners, new studies are required to ensure their biocompatibility and to evaluate their potential health risks²⁰, especially given the aligners’ large surface area. Current evidence shows that 3D-printed aligners may release UDMA²¹, exhibit varying degrees of cytotoxicity depending on the post-curing procedures used²², and have negative effects on human gingival keratinocytes²³. Notably, conventional thermoformed clear aligner resins have also been shown to possess cytotoxic potential²⁴, and elevated levels of inflammatory cytokines (IL-6 and IL-8) in gingival crevicular fluid have been reported following orthodontic treatment with thermoformed clear aligners²⁵.

Given the concerns regarding the biocompatibility of dental monomers, particularly their adverse effects on the immune and endocrine systems, it is crucial to accurately quantify the release of these compounds from both conventional and 3D-printed polymer-based dental materials. However, there is a lack of validated analytical methods specifically designed for the simultaneous detection of BPA and UDMA in material eluates²⁶. To address this gap, we developed and validated a rapid and sensitive liquid chromatography–tandem mass spectrometry (LC-MS/MS) method with polarity switching for their quantification in artificial saliva. This method was subsequently applied to real samples of eluates from two types of materials: (1) a conventional thermoformed resin used for splints and clear aligners and (2) a 3D-printed resin used for splints and palatal plates.

Methods

Chemicals and reagents

UDMA was purchased from BIOSYNTH s.r.o. (Bratislava, Slovakia). Standard of progesterone-d₉ was obtained from Steraloids (Newport, RI, USA). BPA, BPA-d₁₆, diethyl phthalate and diethyl phthalate-d4 were purchased from Sigma-Aldrich (St. Louis, MO, USA). Testosteron-d3 was obtained from Cayman Chemicals (Ann Arbor, MI, USA). Ethyl acetate (≥ 99.9%), n-hexane (≥ 99%), and ammonium fluoride were supplied by VWR International (Wayne, PA, USA). Methanol (≥ 99.9%) and water were from Honeywell Research Chemicals (Charlotte, NC, USA). All solvents and reagents were of LC-MS grade.

Preparation of artificial saliva

The artificial saliva was prepared by dissolving 0.8 g/L NaCl, 1.2 g/L KCl, 0.1 g/L CaCl2·2H2O, 0.3 g/L K2HPO4·3H2O, and 0.1 g/L MgCl2·6H2O in distilled water with pH adjusted to 7.0 in a hospital laboratory (General University Hospital in Prague, Czech Republic)²⁷.

Solution preparation and contamination control

Stock solutions (1 mg/mL) and working solutions (1 µg/mL) in methanol were prepared for both BPA and UDMA. Similarly, stock (1 mg/mL) and working solutions (10 µg/mL) were prepared in methanol for the internal standards BPA-d₁₆ and progesterone-d₉. A calibration mixture was prepared from the individual working solutions, yielding final concentrations of 4 ng/mL for both BPA and UDMA. The internal standard (IS) mixture was prepared in methanol to achieve final concentrations of 100 ng/mL for BPA-d₁₆ and 700 ng/mL for progesterone-d₉. All solutions and mixtures were stored at − 20 °C. Eight-point calibration curves were constructed for BPA and UDMA over the range of 0.016–4 ng/mL.

To minimize potential BPA contamination, all samples and reagents were stored in glass vials and bottles. Furthermore, every step of the protocol was carried out using glass equipment (e.g., Pasteur pipettes, glass beakers, and glass tubes). When preparing the eluates of tested samples, all tools were thoroughly rinsed with methanol prior to use. Each batch included two procedural blanks to monitor contamination^28,29. Overall, BPA contamination was below the limit of detection.

Sample Preparation

For each sample, 10 µL of the internal standard (IS) mixture was pipetted into a glass tube. Subsequently, 0.5 mL of the sample (artificial saliva spiked with a known amount of analytes or artificial saliva eluate from the tested material) was added to the tube, followed by 0.5 mL of saline. The samples were briefly vortexed, and liquid–liquid extraction was performed using 2 mL of a hexane: ethyl acetate mixture (v/v, 3:2) for 1 min. The aqueous phase was then frozen, and the organic phase containing the analytes was decanted into clean tubes and evaporated under vacuum for 45 min at 30 °C. After evaporation, the residue was reconstituted in 100 µL of 30% methanol, vortexed, and 90 µL was transferred to an insert placed inside a vial.

LC-MS/MS conditions

Chromatographic separation was performed at 35 °C with a flow rate of 0.7 mL/min. An ExionLC AD system (Sciex, Concord, Canada) was used for the analysis with a short Kinetex Biphenyl column (50 mm × 3 mm, 2.6 μm; Phenomenex, Torrance, CA, USA) and the corresponding Security Guard ULTRA cartridge system (UHPLC C18 for 3 mm ID columns; Phenomenex, Torrance, CA, USA). Water (A) and methanol (B) were used as mobile phases, with a post-column infusion of 6 mM NH₄F in water at a flow rate of 5 µL/min to improve sensitivity. Gradient elution started at 30% B (0–1.5 min), increased linearly to 98% B (1.5–3.5 min), then to 100% B (3.5–4.6 min), dropped to 30% B (4.60–4.61 min), and was maintained at 30% B from 4.61 min to 6 min.

Detection of the analytes was performed using a QTRAP 6500 + mass spectrometer (Sciex, Concord, Canada) with an electrospray ionization (ESI) probe operating in both negative and positive ionization modes. The optimized conditions for UDMA and progesterone-d₉ were as follows: ion spray voltage of 5500 V, temperature of 500 °C, curtain gas at 35.0 psi, ion source gas 1 at 60.0 psi, and ion source gas 2 at 50.0 psi. For BPA and BPA-d₁₆, the ion spray voltage was − 4500 V, with the same temperature (500 °C), curtain gas at 35.0 psi, ion source gas 1 at 40.0 psi, and ion source gas 2 at 50.0 psi. The system was controlled using Analyst 1.7 software. Data were evaluated using Sciex OS software version 2.1.6.

Method validation

The method was validated according to FDA³⁰ guidelines. It was developed and optimized in terms following bioanalytical parameters: selectivity and specificity, sensitivity, accuracy, precision and recovery, stability and matrix effects (ME).

Selectivity and specificity were assessed using blank artificial saliva from two sources, where potential interferences in retention times of analytes and IS were looked for. MEs and extraction recovery (RE) are associated with these parameters, assessing potential ion suppression or enhancement, as well as the efficiency of the extraction process. While the artificial saliva contained trace amounts of BPA, MEs were evaluated by comparing slopes of regression lines in 0.9% saline and artificial saliva similarly as Higashi et al.³¹. REs were performed according to Matuszewski et al.³² using following formula:

Sensitivity was defined by the lower limit of quantification (LLOQ), determined as five times the analyte signal observed in the zero calibrator (5 × signal-to-noise ratio, S/N), in accordance with FDA and EMA Guidelines. The limit of detection was set at 3 x the S/N ratio.

Accuracy and precision (A&P) were assessed in three independent A&P runs at four different concentrations: (I) artificial saliva, (II) artificial saliva with the addition of 0.08 ng/mL of BPA and UDMA, (III) artificial saliva with 0.4 ng/mL of BPA and UDMA, and (IV) artificial saliva with 0.96 ng/mL of BPA and UDMA. Within-run and between-run precision were expressed as the coefficient of variation (CV), which should be within ± 15%, except at the LLOQ concentration, where a ± 20% variation is acceptable. The method’s accuracy was determined using the following formula:

where c(A) is the mean analyte concentration in the spiked sample, c(B) is the mean concentration in the non-spiked sample, and c(C) spiked concentration.

Following stability tests were performed: (1) bench-top stability to determine the stability of samples in laboratory temperature; (2) processed sample stability to evaluate residence time in autosampler; (3) long-term stability of the samples to evaluate the storage temperature (2–8 °C) of the samples during the whole protocol.

For the bench-top stability test, six replicates of low QC samples (spiked with 0.08 ng/mL of BPA and UDMA) and six replicates of high QC samples (spiked with 0.96 ng/mL of BPA and UDMA) in artificial saliva were prepared. The samples were kept at room temperature for 4 h and compared with freshly prepared QCs and calibrators. The groups were compared using the Mann–Whitney test.

For the processed sample stability test, six replicates of low QC samples (spiked with 0.08 ng/mL of BPA and UDMA) and six replicates of high QC samples (spiked with 0.96 ng/mL of BPA and UDMA) in artificial saliva were prepared and analyzed. Subsequently, this set was reanalyzed after 24 h together with freshly prepared calibrators. The comparison was performed using the Wilcoxon test for paired samples.

The long-term stability test was performed on 12 artificial saliva eluates stored in a refrigerator (2–8 °C) for three months. The concentrations measured at the beginning and after three months were compared using the Wilcoxon test for paired samples.

Freeze–thaw stability testing was not performed, as the samples were not frozen at any stage and freezing of future samples is not anticipated.

Tested materials

We utilized the newly developed method to quantify BPA and UDMA in eluates from two different dental polymer-based materials. DURAN+ (Scheu-Dental, Iserlohn, Germany), a thermoformed three-layer material used for splints and clear aligners, was selected because it was previously shown to release UDMA³³. According to the safety data sheet, DURAN + is made of polyethylene terephthalate glycol (PETG) and has a removable polyethylene (PE) protective film. KeySplint Soft (Keystone Industries GmbH, Singen, Germany), a material used for 3D printing splints, night guards, and bleaching trays, was selected as a negative control, because it was expected to contain neither UDMA nor BPA based on the safety data sheet. However, we also aimed to verify that BPA would not be released, because of a novel intended application as palatal plates in children with a cleft palate. KeySplint Soft contains 2-phenoxyethyl methacrylate, isobornyl methacrylate, HEMA, and trimethylbenzoyl diphenylphosphine oxide (TPO).

Five specimens of each material were fabricated. DURAN + aligners were vacuum-formed on a 3D-printed model at 220 °C and a pressure of 6.2 bar (Biostar; Scheu-Dental, Iserlohn, Germany), and specimens corresponding to a lower premolar were cut out of the aligner. For KeySplint Soft, specimens were 3D-printed as cuboids (approximately 10 ⋅ 5 ⋅ 2.5 mm³) with rounded edges using a direct light processing (DLP) 3D printer (LD-002R; Creality, Shenzhen, China) with a wavelength range 385–405 nm. The product-specific and gate-defined settings were used. After printing, the specimens were placed in an isopropanol bath, post-cured, finished, and polished. While the different design and manufacturing of the specimens could influence kinetics of the release, we aimed to mimic the use case rather than to compare the selected materials.

The specimens were weighed, placed into glass tubes with 2 mL of artificial saliva and incubated at 37 °C for 1 day. After 1 day, the eluates were collected and replaced with 2 mL of fresh artificial saliva, then incubated for an additional 6 days with daily manual agitation. Afterwards, the eluates were collected, and 500 µL of each eluate was used for the LC-MS/MS analysis.

Statistical analysis

For statistical evaluation, data below the LLOQ were substituted with LLOQ/√2, in accordance with conventional practice³⁴. Descriptive statistics, including means and standard deviations, were calculated using Microsoft Excel 2024 Version 2408 (Microsoft Corp., Redmond, WA, USA). Because the stability testing data did not follow a Gaussian distribution, non-parametric alternatives to paired t-test and t-test for independent samples were used: the Wilcoxon test for paired samples and the Mann–Whitney test for independent samples. Statistical analyses were performed using Statgraphics Centurion XVI version 16.2.04 (Statpoint Inc., Warrenton, VA, USA; https://www.statgraphics.com) and MedCalc software version 20.210 (MedCalc Software Ltd., Ostend, Belgium; https://www.medcalc.org).

Results

Analytical method parameters

The retention times, precursor ions, fragment ions and MS optimized conditions (declustering potential, DP; entrance potential, EP; collision entrance potential, CEP; collision energy, CE; collision cell exit potential, CXP) for all analytes are given in Table 1.

Table 1.

Retention times, precursor ions, fragment ions, MS Polarity and MS optimized conditions (declustering potential, DP; entrance potential, EP; collision energy, CE; collision cell exit potential, CXP) for all analytes.

Analyte	Retention time (min)	Precursor ion	Type of ion	Quantification ion	Confirmation ion	Polarity	DP (V)	EP (V)	CE (V)a	CXP (V)a
BPA	3.23	227.079	[M - H]−	212.049	133.043	negative	−30	−10	−24 (−32)	−17 (−15)
BPA-d16	3.21	241.088	[M - H]−	222.989	141.877	negative	−45	−10	−28 (−32)	−23 (−17)
UDMA	3.68	493.221	[M + Na]+	407.235	321.199	positive	1	10	37 (45)	22 (18)
Progesterone-d9	4.21	324.403	[M + H]+	128.889	100.055	positive	91	10	23 (27)	14 (12)

^a Values for the confirmation ion are given in the brackets.

Representative chromatograms of UDMA and BPA in artificial saliva are shown in Fig. 2.

Fig. 2.

Representative chromatograms of UDMA and BPA in artificial saliva. Measured concentrations were 0.611 ng/mL for UDMA and 0.814 ng/mL for BPA.

Selectivity, specificity and carry-over

Blank and zero calibrator were free of interference in the retention time of UDMA. In the case of BPA, there was a small peak of BPA which was not quantifiable. Additionally, IS response in the blank did not exceed 5% of the average IS responses of the calibrators. There were no cross-reacting molecules, while artificial saliva contains just few chemicals. We evaluated carry over in blank sample after injection of the highest calibrator. There was no carry-over neither for UDMA nor for BPA.

Sensitivity

According to FDA guidelines, we calculated LLOQ for BPA and UDMA as ≥ five times the analyte response of the zero calibrator in the artificial saliva. The LLOQ for BPA and UDMA are 15 pg/tube (equivalent to 30 pg/mL or 0.13 nmol/L) and 5 pg/tube (equivalent to 10 pg/mL or 0.022 nmol/L), respectively. In accordance, in these concentrations, the coefficients of variation (CVs) for accuracy and precision were within ± 20% (calculated from 6 replicates in three runs).

Accuracy and precision (A&P)

Three independent accuracy and precision (A&P) experiments were conducted at four concentration levels (blank artificial saliva, low, medium and high-quality control (QC), corresponding to addition 0, 0.08, 0.4 and 0.96 ng/mL for both analytes. The final parameters of these runs are shown in Table 2.

Table 2.

Parameters for accuracy and precision runs for BPA and UDMA.

Analyte	Added (ng/ml)	Precision (%)		Accuracy (%)
		Within-run	Between-run
BPA	0.08	3.9	15.0	102.1
	0.4	6.8	7.1	113.9
	0.96	7.2	1.8	105.4
UDMA	0.08	9.7	13.9	113.5
	0.4	7.5	7.7	88.6
	0.96	14.3	6.5	90.7

Recovery and matrix effects

Extraction recovery (RE) in low and high QC samples and slopes of regression lines in artificial saliva and in 0.9% saline with corresponding CVs and coefficients of determination are shown in Table 3. The slopes of regression lines did not significantly differ between artificial saliva and 0.9% saline. This result demonstrates that the artificial saliva matrix had no impact on the determination of BPA and UDMA. Furthermore, the use of internal standards minimizes the impact of matrix in real samples.

Table 3.

Extraction recovery (RE) and matrix effects for BPA and UDMA evaluated by comparing slopes of a regression lines in 0.9% saline and the artificial saliva.

Analyte	RE (%)	RE (%)	Calibration curve in 0.9% saline			Calibration curve in artificial saliva			p-value
	Low QC	High QC	Slope	CV	r2	Slope	CV	r2
BPA	115.1	109.6	0.9201	8.28	0.994	0.9121	13.20	0.991	0.701
UDMA	87.7	93.2	0.0267	9.60	0.988	0.0263	14.83	0.982	0.689

CV - coefficient of variation, r2 – coefficient of determination.

Stability tests

The bench-top stability test (samples at laboratory temperature for 4 h), processed sample stability test (24 h in the autosampler), and long-term stability of the eluates stored at 2–8 °C for 3 months were evaluated. The results of the statistical tests, expressed as p-values, are shown in Table 4.

Table 4.

Stability tests. Differences between groups are expressed by p-values of the test. The significant differences are depicted in bold.

Type of stability test		BPA p-value	UDMA p-value	Performed statistical test
Bench-top	Low QC	0.054	0.004	Mann-Whitney test for independent samples
	High QC	0.093	0.026	Mann-Whitney test for independent samples
Processed sample		0.007	0.002	Wilcoxon test for paired samples
Long-term		0.286	0.176	Wilcoxon test for paired samples

The results show that UDMA is not stable when samples are left at room temperature for 4 h. Therefore, samples must be processed immediately or stored on ice. Furthermore, differences in BPA and UDMA concentrations were observed after samples were left in the autosampler for 24 h (processed sample stability test), indicating that the residence time in the autosampler should be shortened. Additional testing showed that 10 h in the autosampler did not affect BPA and UDMA concentrations. Importantly, when eluates are stored at cold temperatures (2–8 °C), the samples remain stable for at least 3 months.

Comparison of tested materials

Concentrations of UDMA and BPA in the 1-day and 1-week eluates of the tested thermoformed and 3D-printed materials are shown in Table 5. After 1 day, concentrations of both BPA and UDMA were low but measurable in all eluates of the thermoformed DURAN+, while they were below LLOQ for most eluates of the 3D-printed KeySplint Soft. After 1 week, all eluates exhibited concentrations below LLOQ.

Table 5.

Concentrations of BPA and UDMA in eluates of the tested thermoformed and 3D-printed material after 1 day and 1 week. Values lower than LLOQ were replaced by LLOQ/√2, i.e. 21.2 pg/mL for BPA and 7.1 pg/mL for UDMA.

Elution time	Sample	Thermoformed material		3D-printed material
		BPA (pg/mL)	UDMA (pg/mL)	BPA (pg/mL)	UDMA (pg/mL)
1 day	1	114	167	< LLOQ	< LLOQ
	2	152	115	59	12
	3	88	151	31	< LLOQ
	4	45	165	< LLOQ	< LLOQ
	5	374	176	< LLOQ	< LLOQ
	Mean ± SD	154.6 ± 128.7	154.8 ± 24.0	30.7 ± 16.4	8.1 ± 2.2
1 week	1	< LLOQ	< LLOQ	< LLOQ	< LLOQ
	2	< LLOQ	< LLOQ	< LLOQ	< LLOQ
	3	< LLOQ	< LLOQ	< LLOQ	< LLOQ
	4	< LLOQ	< LLOQ	< LLOQ	< LLOQ
	5	< LLOQ	< LLOQ	< LLOQ	< LLOQ
	Mean ± SD	< LLOQ (21.2)	< LLOQ (7.1)	< LLOQ (21.2)	< LLOQ (7.1)

Subsequently, the results were expressed relative to specimen weight (in pg/g), Table 6. The mean weights of thermoformed and 3D-printed specimens were 0.0849 ± 0.006 g and 0.1801 ± 0.007 g, respectively. The amounts of both UDMA and BPA released from the thermoformed DURAN + after 1 day were substantially higher concentrations compared to the 3D-printed KeySplint Soft, and a high standard deviation was observed for BPA. After 1 week, the amounts of released BPA and UDMA were also higher for the thermoformed material, but since all these eluates had concentrations lower than LLOQ, this is only caused by the different weight of the specimens.

Table 6.

Release of BPA and UDMA relative to specimen weight (pg/g) expressed as mean ± SD. The mean weights of thermoformed and 3D-printed specimens were 0.0849 ± 0.006 g and 0.1801 ± 0.007 g, respectively.

Elution time	Thermoformed material		3D-printed material
	BPA (pg/g)	UDMA (pg/g)	BPA (pg/g)	UDMA (pg/g)
1 day	3731 ± 3236	3673 ± 713	338 ± 162	88 ± 18
1 week	517 ± 26	167 ± 12	269 ± 37	84 ± 10

Discussion

In this study, we developed and validated a fast method for analyzing BPA and UDMA in artificial saliva with low LLOQs, suitable for assessing elution from dental polymer-based materials.

All validation experiments for BPA and UDMA yielded satisfactory results. However, careful handling of the samples is essential, as the stability of the analytes—particularly UDMA—is limited due to the high reactivity of its methacrylate groups. During sample preparation, the samples must not be left at room temperature for an extended period; they should be processed immediately or kept on ice. Similarly, processed samples should not remain in the autosampler for more than 10 h, as this can lead to inaccurate results. When these precautions are taken into account, the method performs reliably. An important finding is that non-processed samples stored at cold temperatures (2–8 °C) remain stable for at least three months.

Generally, elution in aqueous media (water, artificial saliva, 0.9% saline) is less pronounced than in organic solvents³⁵, which represent the worst-case scenario^2,36,37. We opted to elute the materials in artificial saliva, as it is most representative of endogenous saliva. Elution in methanol was also attempted; however, it led to the decomposition of 3D-printed specimens which precluded reliable analyses. A comparison of the slopes of the calibration curves in artificial saliva and 0.9% saline shows that the matrix has no significant effect. Therefore, samples can also be eluted in 0.9% saline with comparable results. An advantage of using 0.9% saline is that it is a simpler matrix and thus has a potentially lower risk of BPA contamination (trace amounts of BPA were detected in artificial saliva).

LLOQs for UDMA and BPA were very low in our study; the LLOQ (30pg/mL = 0.13 nmol/L) for BPA in artificial saliva was the lowest in our study in comparison with other studies: 100pg/mL³⁸, 250pg/mL³⁹, 15ng/mL⁴⁰, 500ng/mL⁴¹, [0.3ppm = 0.3 mg/L = 300ng/mL⁴², 230ng/mL⁴³, 50ng/mL³⁶, 90pg/mL in water⁴⁴, 250ng/mL²¹ and 500ng/mL⁴⁵. Similarly, the LLOQ of UDMA in our study (10 pg/mL = 0.022 nmol/L) was the lowest compared to other studies: 1.5ng/mL⁴⁰, 5ng/mL³⁶, 1 µg/mL⁴⁵, 250ng/mL²¹ or 0.31 nmol/L³⁵. The difference in LLOQs may be due to different methodological approaches or calculations; however, increasingly sensitive instrumentation continues to improve detection capabilities. Furthermore, since BPA ionizes poorly, a derivatization step is useful for enhancing sensitivity^28,39. Similar sensitivity can be achieved by using a suitable additive; for example, ammonium fluoride provided comparable sensitivity to derivatization with dansyl chloride⁴⁶.

Naturally, the use of internal standards is highly recommended for each analysis to compensate for losses during sample preparation and to correct for matrix effects. The deuterium-labeled standard for BPA is commercially available, however, the deuterium-labeled standard for UDMA is not. Therefore, we evaluated several internal standards that we had in our possession, including progesterone-d9, testosterone-d3, along with diethyl phthalate²⁶ and diethyl phthalate-d₄³⁵, which have previously been employed by other research groups. The internal standard was selected based on the evaluation of accuracy and precision at three concentration levels and extraction recovery at two levels, with particular emphasis on achieving the lowest possible coefficient of variation. Among the tested candidates, progesterone-d9 demonstrated the most reliable correction and the lowest variability and was therefore selected as the internal standard for UDMA.

To evaluate the analytical method on real samples, we tested the elution of BPA and UDMA from two dental polymer-based materials. The thermoformed material (DURAN+) released more BPA and UDMA than the 3D-printed resin (KeySplint Soft), but the amounts were very low even for DURAN+. Moreover, the materials must be compared with caution, as they were processed using different methods and specimens differed in shape and size, which can influence monomer release². The release of UDMA from DURAN + was reported in a previous study³³, where it was paired with a UDMA-containing material Durasplint LC (Scheu-Dental, Iserlohn, Germany). Since Durasplint LC was not used in this study and a protective film was used to prevent contamination from the 3D-printed model during vacuum-forming, we speculate that the minute amounts of UDMA and BPA quantified after 1 day could be a result of contamination from the laboratory where specimens were fabricated.

Assuming that the weight of a splint/aligner for upper/lower dental arch would be approximately 2 g (i.e., 4 g for both arches), the mean amount of BPA released from the thermoformed material after 1 day (< 4 ng/g) would equal the strict TDI of 0.2 ng/kg bw proposed by EFSA¹⁷ for a person weighing 75 kg. While this calculation shows that TDI would be exceeded for adolescents (who weigh less and often receive orthodontic treatment), this would be only true on the first day of using an aligner, as the values in eluates collected after 1 week, reflecting release on days 2–7, were below LLOQ. The decrease in monomer release after the first day has been consistently reported by other studies of dental materials as well^47–50. On the other hand, new aligners are introduced regularly (every 1–2 weeks) for extended periods of time, repeatedly exceeding TDI. Furthermore, mechanical stress inflicted on the aligners upon chewing, changes of temperature and pH, as well as other factors may affect the kinetics of BPA and UDMA release, possibly leading to higher exposures. Finally, it is noteworthy that a high degree of variation in BPA release was observed for the thermoformed material. While higher deviation is common in values close to LLOQ (30 pg/mL for BPA), the within-run precision for BPA was below 7.2% in the validation experiments. Therefore, the higher variability suggests that the material may be inconsistent and/or that its handling was not perfectly standardized.

The amounts of both BPA and UDMA released from most 3D-printed specimens after 1 day and all 3D-printed specimens after 1 week were below the respective LLOQs. Therefore, the release of these monomers from the tested 3D-printed material (KeySplint Soft) has not been definitively demonstrated, and from the perspective of BPA and UDMA release, the use of KeySplint Soft appears to be safe even for palatal plates used in infants. Nevertheless, the results cannot guarantee the overall safety of the material, as potential adverse effects could be associated with other ingredients of the material listed in the safety data sheet – 2-phenoxyethyl methacrylate, isobornyl methacrylate, HEMA, and TPO. In summary, while additional investigations are needed to comprehensively assess and compare the materials or processing methods (thermoforming versus 3D-printing), the results demonstrate that the method is reliable for testing different types of dental polymer-based materials even if very low amounts of BPA and UDMA are released.

This rapid LC-MS/MS method enables a reliable and sensitive analysis of BPA and UDMA release from dental polymer-based materials in artificial saliva or other aqueous media. Considering ongoing concerns about the biocompatibility of dental monomers, such methods are necessary to evaluate the potential adverse effects of these analytes on the immune and endocrine systems. In light of EFSA’s proposed reduction of the TDI for BPA, a low LLOQ is particularly important.

Author contributions

Jana Vitku: Writing – original draft, Conceptualization, Supervision, Funding acquisition; Tereza Skodova: Writing – review & editing, Validation, Methodology; Ye Eun Tak: Writing – review & editing, Investigation, Methodology; Babak Sayahpour: Writing – review & editing, Investigation; Sarah Bühling: Writing – review & editing, Investigation; Tereza Srolerova: Writing – review & editing, Methodology; Lucie Kolatorova: Writing – review & editing, Validation, Anezka Varausova: Writing – review & editing, Methodology; Antonin Tichy: Writing – original draft, Conceptualization, Supervision, Project administration.

Funding

This work was supported by MH CZ—DRO (Institute of Endocrinology - EÚ, 00023761) and by Charles University, project Cooperatio 207030 Dental Medicine/LF1.

Data availability

The datasets generated during the current study are available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.